You need to agree to share your contact information to access this dataset
This repository is publicly accessible, but you have to accept the conditions to access its files and content.
WorldSimReady-Home is available under CC BY-NC-SA 4.0. You must comply with its Attribution, NonCommercial, and ShareAlike terms. By submitting this form, you agree to share your email address, username, and the information below with the dataset authors. Access approval does not grant commercial-use rights.
Log in or Sign Up to review the conditions and access this dataset content.
WorldSimReady-Home
Dataset description
|
CAD-based SimReady assets |
Manually reviewed scenes |
Scalable task generation |
WorldSimReady-Home is built from CAD-based object assets, optimized and enriched with physical properties to create simulation-ready assets. The assets are assembled into household scenes, each manually reviewed for correct physics configuration.
A scalable simulation task data generation framework builds on these scenes to generate data across robot embodiments and tasks.
Scene dataset statistics
| Metric | Scale |
|---|---|
| Household scenes | 500 |
| Placed object entries | 31,711 |
| Rooms | 4,462 |
| Total floor area | 36,601.61 mΒ² |
Task dataset statistics
| Metric | Scale |
|---|---|
| Robot trajectories | 1,000 episodes |
| Recorded frames | 1,316,359 |
| Recorded duration | 12.18 hours |
| Recording frequency | 30 Hz |
| RGB-D resolution | 640 Γ 480 |
Visual overview
Household scenes
Explore Diverse Home Layouts
Explore RGB / Depth / Semantic
| RGB | Depth | Semantic |
|---|---|---|
Object assets
Explore Diverse Object Assets
Our assets include visual materials and configured physical properties for simulation.
| Asset feature | Included properties |
|---|---|
| CAD-based geometry | Object geometry, transforms, and bounding boxes. |
| Visual materials | Textures, roughness, and metallic parameters. |
| Physical properties | Mass, density, static and dynamic friction, and restitution. |
Robot tasks and recorded observations
Explore Recorded Observations
Dataset structure
Scene packages contain household environments and their object assets. Scene metadata describes the scene and its objects. Trajectory files store robot observations, states and action targets over time.
Scene packages
WorldSimReady-Home/
βββ .gitattributes
βββ README.md
βββ scene/
β βββ scene000000/
β β βββ scene000000.usdz
β β βββ scene000000_metadata.json
β β βββ scene000000.png
β βββ ...
βββ task/ ...
| File | Contents |
|---|---|
<scene_id>.usdz |
USD scene package. |
<scene_id>_metadata.json |
Scene metadata, including object and material records. |
<scene_id>.png |
Overhead scene preview. |
Task recordings
WorldSimReady-Home/
βββ .gitattributes
βββ README.md
βββ scene/ ...
βββ task/
βββ sim/
βββ pick_and_place_tasks/
βββ g1_omnipicker/
βββ pick_and_place_on_the_table/
βββ data/
β βββ chunk-000/
β βββ episode_000000/
β β βββ dataset-driven_object_grasping_and_placement.hdf5
β β βββ meta/
β β βββ info.json
β β βββ episodes.jsonl
β β βββ tasks.jsonl
β β βββ recording_summary.json
β βββ ...
βββ meta/
βββ info.json
βββ episodes.jsonl
βββ tasks.jsonl
βββ manifest.jsonl
The meta/ directory alongside data/ stores dataset-level metadata and indexes. Each episode directory shown in the example contains an HDF5 trajectory file and a separate meta/ directory storing metadata for that trajectory.
HDF5 files store observations, actions and episode flags, with metadata in JSON / JSONL files. Our task dataset format is based on the RLDS format used by Open X-Embodiment.
Usage
Download the scene dataset
hf download Yootta/WorldSimReady-Home --repo-type dataset \
--include "scene/*" --local-dir worldsimready-home
Download the task dataset
hf download Yootta/WorldSimReady-Home --repo-type dataset \
--include "task/*" --local-dir worldsimready-home
Contributors
Shangpin Home Collection and Yootta are joint contributors to WorldSimReady-Home, supporting the development and release of the dataset.
License and citation
License: CC BY-NC-SA 4.0 , following the repository declaration.
If you use WorldSimReady-Home in your work, please cite the dataset:
@misc{world_simready_home,
author = {{Yootta}},
title = {{WorldSimReady-Home}},
year = {2026},
howpublished = {\url{https://huggingface.co/datasets/Yootta/WorldSimReady-Home}}
}
- Downloads last month
- 75